Differences

This shows you the differences between two versions of the page.

--- tamiwiki:projects:egpu [2023/06/18 18:07] – [P40] yair
+++ tamiwiki:projects:egpu [2023/11/04 11:07] (current) – [1080Ti] yair
@@ Line 1: / Line 1: @@
 ====== EGPU ======
-https://docs.kernel.org/admin-guide/thunderbolt.html
+{{ :tamiwiki:projects:pasted:20230618-183833.png}}
+we are using the [[https://egpu.io/best-egpu-buyers-guide/|TH3P4G3 eGPU external thunderbolt]] thing.\\
+Linux Kernal notes > https://docs.kernel.org/admin-guide/thunderbolt.html\\
+[[https://realtechtalk.com/Nvidia_Tesla_GPUs_K40K80M40P40P100V100_at_homedesktop_hacking_cooling_powering_cable_solutions_Tutorial_AIO_Solutions-2465-articles|realtechtalk guide]], [[https://archive.is/Kgj7E|mirror]]
+=== ThunderBolt check and setup ===
 TLDR
   - upgrade kernel (??)
@@ Line 46: / Line 55: @@
 <code bash>
 (base) user@eight:~$ echo 1 | sudo tee /sys/bus/thunderbolt/devices/0-1/authorized
-(base) user@eight:~$ sudo ubuntu-drivers devices
-== /sys/devices/pci0000:00/0000:00:1c.4/0000:04:00.0/0000:05:01.0/0000:07:00.0/0000:08:01.0/0000:09:00.0 ==
-modalias : pci:v000010DEd00000DD8sv000010DEsd0000084Abc03sc00i00
-vendor   : NVIDIA Corporation
-model    : GF106GL [Quadro 2000]
-manual_install: True
-driver   : nvidia-driver-390 - distro non-free recommended
-driver   : xserver-xorg-video-nouveau - distro free builtin
 </code>
-just an old card...
-but EEK
-<code bash>
-(base) user@eight:~$ lspci | tail
-:04.0 PCI bridge: Intel Corporation JHL7540 Thunderbolt 3 Bridge [Titan Ridge DD 2018] (rev 06)
-:00.0 VGA compatible controller: NVIDIA Corporation GF106GL [Quadro 2000] (rev a1)
-:00.1 Audio device: NVIDIA Corporation GF106 High Definition Audio Controller (rev a1)
-$sudo dmesg
-[ 1041.053826] nvidia: module license 'NVIDIA' taints kernel.
-[ 1041.053831] Disabling lock debugging due to kernel taint
-[ 1041.484017] nvidia-nvlink: Nvlink Core is being initialized, major device number 509
-[ 1041.484032] NVRM: The NVIDIA Quadro 2000 GPU installed in this system is
-               NVRM:  supported through the NVIDIA 390.xx Legacy drivers. Please
-               NVRM:  visit http://www.nvidia.com/object/unix.html for more
-               NVRM:  information.  The 535.43.02 NVIDIA driver will ignore
-               NVRM:  this GPU.  Continuing probe...
-[ 1041.501047] NVRM: No NVIDIA GPU found.
-[ 1041.521176] nvidia-nvlink: Unregistered Nvlink Core, major device number 509
-[ 1042.332830] nvidia-nvlink: Nvlink Core is being initialized, major device number 509
-[ 1042.332842] NVRM: The NVIDIA Quadro 2000 GPU installed in this system is
-               NVRM:  supported through the NVIDIA 390.xx Legacy drivers. Please
-               NVRM:  visit http://www.nvidia.com/object/unix.html for more
-               NVRM:  information.  The 535.43.02 NVIDIA driver will ignore
-               NVRM:  this GPU.  Continuing probe...
-[ 1042.335282] NVRM: No NVIDIA GPU found.
-[ 1042.335835] nvidia-nvlink: Unregistered Nvlink Core, major device number 509
-</code>
-<WRAP center round alert 33%>WE ARE TAINTED</WRAP>
-==== driver ====
-we went with ubuntu selection
-but cute https://www.nvidia.com/en-us/drivers/unix/
-<code bash>
-$ sudo apt installl nvidia-headless-535
-#downgrade nvidia to quadro supported version
-sudo apt install nvidia-headless-390
-# EEK
-RROR (dkms apport): kernel package linux-headers-6.3.7-060307-generic is not supported
-Error! Bad return status for module build on kernel: 6.3.7-060307-generic (x86_64)
-Consult /var/lib/dkms/nvidia/390.157/build/make.log for more information.
-dpkg: error processing package nvidia-dkms-390 (--configure):
- installed nvidia-dkms-390 package post-installation script subprocess returned error exit status 10
-dpkg: dependency problems prevent configuration of nvidia-headless-390:
- nvidia-headless-390 depends on nvidia-dkms-390; however:
-  Package nvidia-dkms-390 is not configured yet.
-dpkg: error processing package nvidia-headless-390 (--configure):
- dependency problems - leaving unconfigured
-Processing triggers for libc-bin (2.36-0ubuntu4) ...
-No apport report written because the error message indicates its a followup error from a previous failure.
-                                                                                                          /sbin/ldconfig.real: /lib/lib
-ndi.so.4 is not a symbolic link
-Processing triggers for man-db (2.10.2-2) ...
-Processing triggers for initramfs-tools (0.140ubuntu17) ...
-update-initramfs: Generating /boot/initrd.img-6.3.7-060307-generic
-Errors were encountered while processing:
- nvidia-dkms-390
- nvidia-headless-390
-</code>
-downgrading but to headless,\\
-without touching the x config?
-going with [[https://github.com/Avalon-Benchmark/avalon|avalon]] readme
-[1]  A 3D video game environment and benchmark designed from scratch for reinforcement learning research
-<code bash>
-conda create -n avalon python=3.9
-conda activate avalon
-sudo apt install --no-install-recommends libegl-dev libglew-dev libglfw3-dev libnvidia-gl libopengl-dev libosmesa6 mesa-utils-extra
-#this will also install torch...
-pip install avalon-rl[train]
-python -m avalon.install_godot_binary
-python -m avalon.common.check_install
-</code>
-why even bother, the quaDRO IS JUST A TEST. \\
-NEED TO CLEAN REMOVE THE 390 driver AND MOVE BACK TO
-NVIDIA-CURRENT
-==== P40 ====
-NVIDIA Tesla P40 24GB DDR5 GPU Accelerator Card Dual PCI-E 3.0 x16\\
-need to retrofit with a FAN,it doesnt come with one
-got one on ebay for 200$(+shipping) ([[https://archive.md/SL4Kq|ebay mirror]])\\
-some dude got it working, https://github.com/JingShing/How-to-use-tesla-p40
-{{ :tamiwiki:projects:pxl_20230615_095016755.jpg?400 |}}
-installing
-<code bash>
-sudo apt install nvidia-headless-535
-</code>
-there is some issue, unlike other cards the blue led doesnt turn green on thunderbolt connection.\\
-no power passing to the gPU.\\
-unlike with other cards we tried (quadro 2000 and 660Ti)
-:(
 ==== 1080Ti ====
@@ Line 223: / Line 104: @@
 $ sudo ubuntu-drivers autoinstall
-The following additional packages will be installed:
-  libnvidia-common-535 libnvidia-compute-535:i386
-  libnvidia-decode-535 libnvidia-decode-535:i386
-  libnvidia-encode-535 libnvidia-encode-535:i386
-  libnvidia-extra-535 libnvidia-fbc1-535 libnvidia-fbc1-535:i386
-  libnvidia-gl-535 libnvidia-gl-535:i386 nvidia-prime
-  nvidia-settings nvidia-utils-535 screen-resolution-extra
-  xserver-xorg-video-nvidia-535
-The following packages will be REMOVED:
-  libnvidia-common-390 libnvidia-gl-390
-The following NEW packages will be installed:
-  libnvidia-common-535 libnvidia-compute-535:i386
-  libnvidia-decode-535 libnvidia-decode-535:i386
-  libnvidia-encode-535 libnvidia-encode-535:i386
-  libnvidia-extra-535 libnvidia-fbc1-535 libnvidia-fbc1-535:i386
-  libnvidia-gl-535 libnvidia-gl-535:i386 nvidia-driver-535
-  nvidia-prime nvidia-settings nvidia-utils-535
-  screen-resolution-extra xserver-xorg-video-nvidia-535
 </code>
+==== P40 ====
+<WRAP center round important 60%>
+this doesnt work on our test machine
+</WRAP>
+{{ :tamiwiki:projects:pxl_20230615_095016755.jpg?400|}}
+the P40 needs modern motherboard that allow for ''Enable Above 4G memory'' bios see [[https://github.com/JingShing/How-to-use-tesla-p40#bios-settings|link]], see [[tamiwiki:projects:P40a|P40]] page for info on dedicated machine.
+NVIDIA Tesla P40 24GB DDR5 GPU Accelerator Card Dual PCI-E 3.0 x16\\
+need to retrofit with a FAN,it doesnt come with one
+got one on ebay for 200$(+shipping) ([[https://archive.md/SL4Kq|ebay mirror]])\\
+some dude got it working, https://github.com/JingShing/How-to-use-tesla-p40
+=== SPECIFICATIONS: ===
+    * GPU Architecture: NVIDIA Pascal
+    * Single-Precision Performance 12 TeraFLOPS*
+    * Integer Operations (INT8) 47 TOPS* (TeraOperations per Second)
+    * GPU Memory 24 GB
+    * Memory Bandwidth 346 GB/s
+    * System Interface PCI Express 3.0 x16
+    * Form Factor 4.4” H x 10.5” L, Dual Slot, Full Height
+    * Max Power 250 W
+    * Enhanced Programmability with Page Migration Engine Yes
+    * ECC Protection Yes
+    * Server-Optimized for Data Center Deployment Yes
+    * Hardware-Accelerated Video Engine 1x Decode Engine, 2x Encode Engine />
+    * NVPN: 699-2G610-0200-100
+    * NVIDIA® CUDA® cores: 3840
+installing
+<code bash>
+sudo apt install nvidia-headless-535
+</code>
+there is some issue, unlike other cards the blue led doesnt turn green on thunderbolt connection.\\
+no power passing to the gPU.\\
+:(
 ==== misc ====

TAMI - wiki

User Tools

Site Tools

Differences

Page Tools